Download Redesigning the Employment Protection System

Forthcoming, De Economist
Redesigning the Employment
Protection System∗
Olivier Blanchard† and Jean Tirole
‡
December 10, 2003
Introduction
Following repeated attempts to reform Dutch dismissals law in
the past, Professor Rood was asked in 1999 to form a committee to study possible reforms of the dismissals process. The
committee’s report, published in November 2000, called for the
abolition of the preventive check on dismissals by the Center
∗
Paper prepared for the second author’s Tinbergen lecture (October
24, 2003), who thanks the Royal Netherlands Economic Association and
the Dutch Central Bank for their warm hospitality. The paper builds on a
report by the two authors on the reform of employment protection in France
(2003a). The authors are grateful to Frank den Butter, Jan van Ours, and
an editor for helpful comments, and especially to Michèle Belot for both
very detailed information on the Dutch labor market and institutions and
useful reactions to previous drafts.
†
MIT and NBER.
‡
IDEI and GREMAQ (UMR 5604 CNRS), Toulouse, CERAS (URA
2036 CNRS), Paris, and MIT.
Employment Protection Reform
2
for Work and Income and argued in favor of an ex post procedure to resolve disputes. The Committee’s proposals were, at
the request of the government, examined by the bipartite Labor
Foundation (Stichting van de Arbeid, STAR.) The Foundation,
while recognizing weaknesses in the current system, sharply criticized the proposal, thereby stalling further reform. Strikingly,
it took more than two years and a half for the trade unions’
and employers’ organization to come up with a stance on the
report.1
It would be presumptuous for us to assess which side was right —
at best, the arguments developed in this lecture will shed some
light on the relevant trade-offs. Rather, we see this episode as
epitomizing the fact that there may be no labor market institution more controversial than employment protection regulation—
the set of laws and procedures regulating separations between
firms and workers. Even in a country such as the Netherlands,
known for the quality of its social dialog,2 and in a state of low
unemployment, where one would believe the topic could be approached dispassionately, social partners had a hard time reaching a conclusion.
More generally, wherever the issue of employment protection is
1. The proposition to abolish the administrative procedure received strong
support from the largest employers’ association. Trade unions were reluctant, and employers in the small and medium-size business sectors were
keen to keep the existing and relatively cheap procedure.
2. This social dialogue in turn has been facilitated by close networking
between unions, employers’ organizations, the government, and other third
parties (advisory bodies, central bank), as argued in Den Butter and Mosch
2003.
Employment Protection Reform
3
raised:
•
Workers focus on the pain of unemployment, and argue
that such pain should be taken into account by firms
when they consider closing a plant, or laying off a worker.
That workers protected by employment protection would
favor it is no great surprise. But evidence from surveys
shows that support for employment protection is more
general, more broad based.
•
Firms, on the other hand, complain not only about the
direct cost, but also about the complexity and the uncertainty introduced by employment protection regulation.
They argue that it makes it difficult for them to adjust
to changes in technology and product demand, and that
this in turn decreases efficiency, increases labor cost and,
in so doing, deters job creation.
•
Most economists and international economic organizations, from the OECD to the IMF, tend to side with
firms. There is, they argue, a trade-off between insurance
and efficiency. The current systems impede reallocation,
and, by implication, reduce efficiency. They lead to higher
costs, and thus lower employment. At a minimum, they
could and should be made more efficient. More likely,
overall employment protection should be reduced.
Faced with such conflicting demands and advice, the governments of Western Europe have been timid. They have learned,
often the hard way, that workers covered by employment protection are not eager to see it reduced, and that these workers
Employment Protection Reform
4
represent the majority of the labor force, and a large part of the
electorate. So, most recent employment protection reforms have
worked at the margin, through the introduction and extension
of the scope for fixed duration contracts—contracts subject to
more limited employment protection and simpler administrative
rules. For the most part, employment protection for regular contracts has remained unchanged. The evidence so far is that this
dual system has led to an increasingly dual labor market, with
mixed efficiency and negative distributional effects.
We believe that what has been missing most in these discussions
and in these reforms is a clear idea of what “good employment
protection regulation” should look like. Starting from the status
quo, firms and international organizations have argued for less
protection. Workers and unions have fought to keep the protection they had. Governments have looked for politically feasible
incremental reforms. But the ultimate goal, the shape of optimal
employment protection, has been left undefined.
Our goal in this lecture is to explore this question. The lecture
has three parts. We begin by using economic reasoning to draw
the contours of a good employment protection system. We then
describe employment protection systems in France and in the
Netherlands. Having done so, we indicate directions for reform.
Employment Protection Reform
5
1
Economic analysis: letting firms
internalize the social cost of their layoffs
1.1
Broad principle
Let us begin with a broad principle: Economic agents can be
given discretion provided that they bear the cost that their decisions impose on other agents. This “internalization of externalities” ensures that the decision-maker in question reflects others’ preferences and is therefore accountable. For example, it is
widely accepted that polluters ought to pay for the social cost
of their pollution.
The translation of this broad principle in the matter of employment is that firms ought to pay for the cost they impose on
society when they lay workers off. This cost includes the financial cost born by the unemployment insurance (UI) fund and the
psychological and other costs born by the dismissed worker.3 The
former can be addressed through a layoff tax, and the latter, as
it now is, through severance pay. In particular, a full internalization of the impact of a layoff on the UI fund calls for a layoff tax
equal to subsequent unemployment benefits, or, in other words,
for a contribution rate— defined as the ratio of the layoff tax to
unemployment benefits— equal to 1.
The layoff tax could be computed in two ways. It could be a
fixed amount, possibly indexed on various observables such as
age, or location. Alternatively, and following the US experience
rating system put in place in the 1930s, the tax could be levied
3. Collective layoffs may further impose substantial costs on local communities in a depressed area.
Employment Protection Reform
6
“ex post” on the basis of the actual length of unemployment
of the dismissed worker. The latter approach offers two benefits
besides simplicity. First, it better reflects the actual cost of the
layoff for the fund; firms’ layoff decisions therefore react more
finely to the actual re-employability of the worker. Second, ex
post tax assessments incentivize firms to provide better training
to their employees so that they can find a job more quickly after
a layoff.
It may actually be worth describing the “experience rated” systems of unemployment contributions used in the United States
in more detail.4 The systems vary across states, but all base the
payments of firms on their past behavior (as in a bonus-malus
system). The most commonly used system is called the “reserve
ratio” system of unemployment contributions. Leaving aside the
many complicated details, its principle is simple: Each firm has
a running balance with the state unemployment agency, with
contributions by the firm to the fund on one side, and benefits paid by the agency to the workers laid off by the firm on
the other. Once a year, the state computes the net outstanding
balance, and requires the firm to pay some proportion of this
outstanding balance over the following year. The factor of proportionality depends both on the net balance of the firm, and
the net balance of the state fund as a whole. This system has
two implications:
•
Ignoring discounting, and assuming that firms do not go
4. A useful description of the US experience is given in Fougère and Margolis (2000).
Employment Protection Reform
7
bankrupt and do not hit the various ceilings that limit
contributions (all considerations being relevant in practice), firms eventually pay the full cost of unemployment
benefits for the workers they lay off—the contribution
rate is equal to one.
•
The factor of proportionality determines how the timing
of payments depends on current and past layoffs. If the
factor of proportionality is equal to one, so firms are asked
to return to zero balance each year, then payments are
closely related to current (or more precisely last year’s)
layoffs. The lower the factor of proportionality, the more
contributions depend on past layoffs in the more distant
past.
How should one then think about the choice of the factor of proportionality? If firms are operating in a stable, ergodic, environment, going sometimes through good
times, sometimes through bad times, then letting the factor of proportionality be small will make the firm’s contributions depend on its mean observed layoff rate in the
past, which is also equal to the probability of a layoff in
the future. If, however, as is more likely, the underlying
probability changes over time, then a higher factor of proportionality, giving more weight to recent layoffs, will be
closer to the underlying current probability. But it will
impose higher liquidity costs on firms.
Note (and we shall come back to this in more detail) that the
principle of internalization is at odds with the systems of em-
Employment Protection Reform
8
ployment protection prevailing in France and the Netherlands in
at least three related dimensions. First, in these systems, contributions by firms to the unemployment insurance fund take
the form of payroll taxes: A firm with a higher layoff rate does
not pay higher contributions to the unemployment insurance
fund. In other words, the contribution rate is zero. Second, in
the absence of layoff taxes, unemployment benefits must be financed through the payroll tax, providing an added incentive
for firms to lay workers off. Third (and relatedly!), the layoff
process is subject to administrative and judicial control. Firms
have to prove either fault by the worker in the case of an individual layoff, or economic need in the case of collective layoffs. By
contrast, under the economic approach, employment protection
stems from the deterrence effect of a layoff tax rather than from
an administered process.
1.2
Complications
This benchmark of a unit contribution rate defines a useful and
simple guiding principle: a firm should pay to the insurance fund
the full amount of unemployment benefits received by the dismissed worker. In practice, the principle has to be refined. We
now study (separately) four major complications: limited unemployment insurance, firms’ financial fragility, ex post wage
bargaining, and firm and worker heterogeneity, and discuss, in
each case, how they modify our conclusions.5
a) Limited unemployment insurance
5.
See Blanchard-Tirole (2003b) for a formal treatment of these ideas.
Employment Protection Reform
9
Workers should not be, and in practice are not, fully insured
against the unemployment risk: to the extent that the agency
cannot fully monitor the search behavior of the unemployed,
offering anything close to full insurance would lead the unemployed to stop searching and remain unemployed.6
A number of recent reforms of the unemployment system in Europe, such as the PARE in France, have aimed at combining
more generous and longer lasting unemployment benefits with
stronger incentives for the unemployed to accept jobs if offered
by the unemployment agency. These reforms clearly go in the
right direction. They potentially offer better tailored insurance:
If truly no jobs are available, then the unemployed continue to
receive unemployment benefits. And they remove, at least in
principle, some of the problems associated with the open ended
unemployment benefits of the past. But realistically, even the
best designed systems cannot fully eliminate monitoring problems, and so, less than full insurance is optimal.
Under incomplete insurance, the firm exerts two externalities
when laying a worker off: one on the unemployment insurance
fund, and the other on the worker. Because the worker cannot,
for incentive reasons, be fully compensated by the fund, the firm
ought to take the net cost to the worker into account. To lead
6. As with any insurance scheme and as shown theoretically by Pauly
(1974), such incomplete insurance may require prohibiting supplementary
arrangements, as secondary insurers and insurees have incentives to strike
supplementary insurance deals at the detriment of the primary insurer (as
was observed for example in the Netherlands for disability schemes).
For a recent survey of the literature on unemployment insurance, see for
example Frederiksson and Holmlund (2003).
Employment Protection Reform
10
firms to take these costs into account, unemployment contributions by firms to the agency must now exceed the unemployment
benefits paid by the agency to workers: The optimal contribution rate is now greater than one. The lower the feasible level
of insurance, the higher the utility costs that layoffs impose on
laid-off workers and the higher the contribution rate.
Under this deviation from the benchmark, unemployment insurance and employment protection are (imperfect) substitutes.
The poorer the insurance, the higher the optimal degree of employment protection. While the result is normative, this negative
relation appears to be present in the data across Continental
European countries.7 The countries with the highest degree of
employment protection (using the OECD index) are also the
countries where unemployment insurance coverage is relatively
limited.
b) Firms’ financial fragility
Introducing financial fragility considerations raises two distinct
issues: the possibility that a layofff tax may encourage rather
than discourage dismissals and the concern that the tax may be
evaded.
Let us start with the former. In the case of small and medium
enterprises, or else, large undiversified firms, layoffs often occur
when the firm encounters hardship. For such firms, a layoff tax
may compound the difficulties and potentially trigger further
layoffs. Should the contribution rate therefore be brought down
7.
See for example Boeri (2002).
Employment Protection Reform
11
and be less than unitary in order to limit the scope for such
“snowball effects”?
One can entertain one of two opposite views in this respect.
The first is that one should not take the firms’ financial structures as given. A wide array of arrangements across sectors and
companies are observed, that show that financial arrangements
adjust to accommodate the specificities of the companies’ environments. One may therefore expect that firms would in the
long term adjust their financing to reflect the existence of a
layoff tax. They will want to insulate spillovers of an adverse
evolution in their activity A on their otherwise healthy activity B. One may hope that, in response to an increase in layoff
taxes, financial markets will at least partly adjust to alleviate
the problem, providing more funds to the firms in bad times
in order to allow them to pay the now higher layoff tax. Concretely, firms may take on more long-term debt and issue more
equity, obtain larger credit lines from banks, and more generally
alter their risk management. This line of thought presumes that
financial markets will in the long term fully adjust to the new
regulatory environment, and implies that the principle of full
internalization continues to apply.
Alternatively, one can take a less optimistic view of the ability
of firms to adjust their financial structure; furthermore one must
surely do so at the time of the introduction of the layoff tax as
financial structures of existing firms are still run by legacy arrangements. Under that view, in order to avoid snowball effects,
the layoff tax should then be lower than unitary.
Employment Protection Reform
12
To sum up, the firms’ financial fragility implies that the layoff
tax should grow after its initial introduction to reach a level that
is smaller than or equal to the cost of unemployment. How close
the contribution rate should be to one depends on one’s optimism as to the possibility of adjustment of financial structures.
Let us now come to the second issue, the possibility that the
layoff tax be evaded. An analogy with environmental matters
is useful here. There is much evidence that activities that are
environmentally hazardous and therefore subject to contingent
environmental penalties tend to be operated by separate, poorly
capitalized, collateral-free companies. The strategy is one of tax
evasion and consists in leaving the state with a judgment proof
entity in case of pollution or disaster. Although less extreme, layoff taxes may generate qualitatively similar behaviors. Activities
with high layoff risks might be spun off into separate entities;
or firms that encounter hardship may reduce their capitalization or delocalize their new activities abroad in order to evade
future layoff taxes. There is no simple solution to this problem. Fundamentally, the UI fund cannot monitor the solvency
of each and every undertaking in the country. But (imperfect)
solutions exist: the posting of collateral (with the equivalent of
“margin calls” by the UI fund, as the firm’s debit increases); the
requirement of a bank guarantee; liability tracing, that is making “Potentially Responsible Parties” liable as is done in the US
since 1980 in environmental matters through the Comprehensive
Employment Protection Reform
13
Environmental Response, Compensation and Liability Act.
c) Ex post wage bargaining
Other things equal, a transfer of the financing of unemployment benefits from the payroll tax to a layoff tax increases efficiency and therefore encourages job creation. This reasoning,
though, presumes wage moderation or, in the parlance of economics, wage bonding. By contrast, under wage renegotiations,8
the introduction of a layoff tax may raise the cost of labor and
discourage job creation. The reason is that the layoff tax and
the cut in the payroll tax both make the firm more accountable
for its layoffs and thus weaken its bargaining position vis-à-vis
workers.
What should the state then do? It now clearly faces a trade-off:
Making firms pay for the full cost of an additional layoff will
lead firms to take the right decision at the destruction margin:
But this high contribution rate will also increase the bargaining
power of workers, and thus increase the wage. This will increase
the overall cost of labor and will adversely affect job creation.
By how much will depend on the amount of effective bonding.
Conversely, choosing a contribution rate less than one will lead
firms to destroy too many jobs, and lead to too many layoffs. It
will however lead to a smaller increase in the overall cost of labor
and thus have a smaller adverse effect on employment creation.
In short, the more firms are made to pay for the expected cost
of unemployment benefits, the smaller the distortion will be at
8.
In particular if they occur at the firm’s or worker’s level.
Employment Protection Reform
14
the destruction margin, but the larger the distortion will be at
the creation margin.
Thus, to the extent that effective bonding is limited, the contribution rate must now be less than one. It will be closer to one,
the higher the workers’ commitment to wage moderation, or, in
the absence of commitment, the lower their bargaining power.
d) Heterogeneity of firms and workers
Not all firms and all workers are alike. What does a positive
contribution rate imply for their respective fortunes?
Let us first handle firms’ heterogeneity by dividing them into
“weak” and “strong” firms. Firms may be weak for one of several
reasons: they may have lower average productivity; or they may
face more volatile demand; or they may have shallow pockets
and therefore little liquidity to meet adverse shocks. Either way,
such firms will have more layoffs on average and thus will to pay
higher layoff taxes.9 They will therefore have higher labor costs.
This is to some extent as should be, given that they impose
larger social costs. Things are a bit more complex, though, and,
in fact, firm heterogeneity calls for a reduced contribution rate.
Intuitively, if weak firms have an incentive to create jobs, so
do a fortiori strong firms. So addressing the creation margin
requires making it attractive for weak firms to create jobs. By
contrast, at the destruction margin, weak and strong firms face
a similar problem: ex-post, whether to keep the worker given the
9. They will not be able to pass those costs on to workers through lower
wages.
Employment Protection Reform
15
realization of productivity. Put differently, a change in the layoff
tax affects the destruction margin of both weak and strong firms,
but affects the creation margin only for weak firms. A small
reduction in the layoff tax compensated by an increase in the
payroll tax cross-subsidizes weak firms and thereby encourages
job creation. To sum up, firm heterogeneity is a further argument
in favor of a contribution rate below one.10
So is worker heterogeneity, for pretty much the same reason.
“Weaker workers”, those with a lower expected productivity and
a higher layoff probability, or with worse labor market prospects
when unemployed—say because they live in a depressed region—
will need to accept lower wages or else run the risk of remaining
unemployed when the layoff tax is raised. A small reduction in
the layoff tax compensated by an increase in the payroll tax
cross-subsidizes weak workers and therefore raises their employability.
Does worker heterogeneity therefore call for a reduction in the
contribution rate? The answer depends on whether the fragility
of certain categories of workers is observable by the state. Sup10. Interestingly, the Dutch Reduced Pay Fund, which pays for short-term
(less than 33 weeks) benefits to the unemployed, is a sector-level fund (the
General Unemployment Fund, which pays the remaining benefits by contrast is a national fund). Premia are differentiated across sectors (for example, the sector of temporary work agencies had the highest payroll tax
rate: 9.20%). This differentiation is probably good from the point of view
of encouraging job creation in sectors with low layoff rates. On the other
hand, because the fund is financed through a payroll tax rather than a layoff tax, this differentiation does not help with the destruction margin. In
particular, high payroll tax rates in risky sectors raise the firms’ incentive
to lay workers off.
Employment Protection Reform
16
pose first that it is. The “weaker worker” may live in an isolated
labor market or depressed region, be senior, or else registered on
a disability list. In this case, the way to enhance the worker’s
employability is not to reduce the layoff tax (after all, layoffs
of such workers are undesirable), but rather to subsidize the
employer. This subsidy can take the form of a targeted cut in
the payroll tax, as is done for example in the Netherlands for
workers with registered disability. Alternatively, one could design targeted job creation subsidies, although the latter, being
lump sum, are more likely to be abused than payroll tax reductions.
Some other forms of worker fragility cannot however be verified
by the state, at least not easily. In that case, facilitating the
worker’s employability then requires a reduction in the layoff tax,
combined perhaps with the introduction of an unemploymentlength-insensitive component in the structure of the layoff tax,
so as not to penalize too much workers who are likely to stay
unemployed for a long time.
1.3
Other considerations
Two further remarks to close the theoretical analysis:
First, we have focused on layoffs. Separations also occur as workers quit either to another job or into unemployment. In the absence of a layoff tax, the employer and employee have an incentive to reclassify a quit into unemployment as a layoff if, as is
the case in France and the Netherlands, workers who quit do not
receive unemployment benefits. The introduction of a layoff tax
Employment Protection Reform
17
reduces the incentives for such gaming of the insurance system.
In passing, a similar form of gaming used to occur in the Netherlands with respect to entry into disability status (as discussed
in van der Ploeg 2003). Until the reform of 1998, employers
and employees had a common interest in disguising a layoff as
a disability. As there was no premium differentiation between
employers, the latter thereby did not have to incur the cost and
bother associated with dismissals, and workers’ access to the allowance thereby became open-ended. Reforms since 1998 have
reduced the gains from reclassification. Firms have to pay for
part of their former employees’ disability benefits. Furthermore,
there have been reforms and there are further plans to bring
disabled workers back to work, for example by cutting benefits
and by strengthening medical requirements.11
Second, and as we know for example from Belot-van Ours (2002)
and as we have already discussed with the issue of wage moderation, the effect of an institutional reform depends on other
institutional features. Another case in point is the governance
of the unemployment agency. To the extent that the layoff tax
is assessed ex post on the actual benefits paid to the dismissed
workers, employers will have an increased demand for overseeing the agency. This raises the issue of who should have control
rights over the agency. An agency run by employers might hassle
the unemployed into accepting inappropriate jobs. Conversely,
11. The law allows employers to fire disable workers if they have been on a
disability scheme for two years or longer. In some cases, disabled workers can
be fired earlier if, for example, they did not try hard enough to reintegrate
the labor market (this being determined by the UWV, the office in charge
of unemployment benefits).
Employment Protection Reform
18
an agency controlled by the workers or the government (as we
see in France) may exhibit the opposite bias. Independent agencies or else a social-partners cooperative undertaking are the two
most appealing governance structures, but giving them a clear
mission certainly is no easy task.
2
The employment protection system in
France and the Netherlands
Having sketched the contours of an optimal system of employment protection, the next question is how it relates to actual employment protection systems in Europe. In this section, we shall
look at employment protection in France and the Netherlands.
Our purpose is not to give an exhaustive presentation of the two
systems, but rather to present them in such a way as to facilitate
the comparison with the conclusions of the previous section.12
Our tentative conclusions are that, relative to the optimal system, (a) the French system is deeply inefficient and (b) the Dutch
approach exhibits some of the same structural flaws, but their
negative impact has been much alleviated by certain specificities and pragmatic adjustments, which de facto have made the
Dutch labor market more flexible than the French.13
12. Three useful sources on French institutions are Pélissier et al.
(2002)(which presents the legal structure), CFDT (2003) (which gives a
user guide for workers) and JurisClasseur Groupe Lexis-Nexis (2002), which
gives the text and interpretation of the 2002 law, called “Loi de Modernisation Sociale”.
13. On paper, the Netherlands are not that much more flexible than
France, except for the regulation of temporary employment. While em-
Employment Protection Reform
2.1
19
The need for motive
France: The general principle today is the need for motive: The
firm must have and must show “real and serious cause”. Only if
such a cause exists can the firm lay a worker off.
The law distinguishes between two types of layoffs:
•
“Personal” (that is, related to the behavior of the employee.) The firm must show that the layoff is the result
of a “serious misdemeanor” (“faute sérieuse”.) What “serious” means is not clearly defined (One definition, found
in the reference labor law text (Pélissier et al. 2002), is:
“Serious: sufficient to justify the layoff”...). It does not require malicious intent, but it must be more than a “light
misdemeanor” (“faute légère”) which does not justify a
layoff.
•
“Economic” (that is, related to the situation of the firm).
ployment protection legislation indices must be taken with caution, the
following table indicates the analogies in legislation concerning regular
employment in the two countries:
Labour Market Statistics – strictness of employment
protection legislation (EPL)
late 80s
Overall EPL Version 1
Regular employment
Temporary employment
Overall EPL Version 2
Collective dismissals
FRA
2,7
2,3
3,1
late 90s
NLD
2,7
3,1
2,4
FRA
3
2,3
3,6
2,8
2,1
NLD
2,1
3,1
1,2
2,2
2,8
Source: OECD Employment Outlook, 1999
Employment Protection Reform
20
The firm must show that the layoff (or layoffs) are the
result of a “real transformation or elimination of job(s)”.
What this exactly means is even more unclear. The ambiguity, and why this is an issue, is best shown in the recent
case of layoffs at Michelin–Wolber. In July 1999, Michelin decided to lay 451 workers off at its Wolber plant,
at the same time as it was announcing large benefits for
the group as a whole. In February 2002, the labor tribunal concluded that the layoffs were not justified, and
asked Michelin to pay a total of 10 million Euros to the
162 laid off employees who had contested the decision,
or about 60,000 euros per employee. The tribunal argued
that “layoffs for economic reasons cannot be justified on
the basis of improving the competitiveness or the profits
of the firm, but only on the basis of maintaining its competitiveness. In the case of Michelin, the purpose was to
improve competitiveness, and thus the layoffs were not
justified”. (The First Court of Appeals confirmed the decision in october 2003).14 Lest one think that this is an
isolated case, very much the same thinking was embodied
in the December 2001 law, which stated that, only when
other avenues had been exhausted, were layoffs justified.
The legislation provided for only three possible grounds
for economic layoffs: major economic difficulties where
14. A charitable interpretation of the court’s opinion is that the firm should
exercise more restraint with regards to layoffs when it is not liquidity
constrained (such a conditioning would make economic sense). We doubt,
though, that the courts have the ability and the information to make such
business judgments.
Employment Protection Reform
21
all possible solutions have been exhausted, technological
changes endangering the very survival of the company,
reorganization required to ensure the survival of the company. Two of the provisions of the law were subsequently
(January 2002) thrown out by the French Supreme Court
(the Conseil Constitutionel) on the grounds that the law
had moved from the principle that layoffs were justified
if they were required to maintain competitiveness to the
principle that layoffs were justified if they were required
to ensure the survival of the firm—a much more stringent
criterion.
In short, the principles that the courts must use in assessing
whether layoffs are justified are extremely unclear; and certainly
the fact that the firm decided that such layoffs were necessary
is clearly not by itself sufficient proof for the courts.
Netherlands: At a formal level, the Dutch situation does not
appear all that different. The burden of proof in principle also
rests on the employer, who must bring hard evidence that the
employment relationship ought to be terminated. Courts require
a dossier before they are willing to treat cases. For a personal
layoff, the employer must show malfunctioning, disability15 or
serious misbehavior. Economic layoffs are deemed to be justified if “adverse circumstances” call for a reorganization or
a bankruptcy. As in France, the absence of objective criteria
confers substantial discretion upon administrative bodies and
15. See section 1.3 for a description of the process for laying off sick or
disabled workers.
Employment Protection Reform
22
judges.
2.2
Limited direct financial costs
Payroll and layoff taxes
In both countries, payments by firms to cover unemployment
benefits are collected through payroll taxes. In France, the payments come for half from the workers, for half from the firms.
In the Netherlands, there are two funds: One, the “Reduced
Pay Fund”, for short-term unemployed (less than 33 weeks) and
the other, the “General Unemployment Fund”, distributing the
remaining payments. The average payroll tax rates are quite
different for those two funds: the employee plus employer contribution rate is 7.35% for the latter and 1.30% for the former.
Unsurprisingly, in the current, low unemployment, context, the
General Unemployment Fund is running large surpluses. The
Reduced Pay Fund is running a deficit. In both countries, the
rate is independent of the history of layoffs by the firm—in our
terminology, the contribution rate (recall: defined as the ratio of
the layoff tax to unemployment benefits) is zero.
One exception in both countries, relates to the layoffs of older
workers. In France, the “contribution Delalande”, introduced
in 1987, mandates additional payments to the unemployment
agency in case of layoffs of older workers. For large firms (50
employees or more), the contribution is equal to two months for
a 50 year old, increasing to 12 months for a 56 year old, and decreasing back to 6 months for a 59 year old or older (the number
Employment Protection Reform
23
of months being halved for firms with less than 50 employees).
In the Netherlands, workers over 57,5 years are protected directly by a higher severance pay, an exception to the otherwise
prevailing last in–first out rule for selecting laid-off workers, and
protected indirectly by the requirement that the firm pay part
of the unemployment benefit. The share of the unemployment
cost born by the firm remains small, though (from 10% for firms
with less than 6 workers to 30% for firms with over 50 workers).
Such measures go in the direction of discouraging layoffs that
are on average socially costly, but they also give rise to “threshold effects’ as they encourage firms to discharge workers before
they reach the age requirement.
Another exception to the absence of internalization is supplied
by the Dutch sickness and disability system, in which experience
rating was progressively introduced starting in the early 1990s,
following an obvious abuse of the system by employers and employees. Dutch employers now have to pay sick workers full pay
for a year (as opposed to 12 weeks in France). Experience rating
in disability insurance was introduced in 1998. Today, employers
have incentives not to have their employees disabled: They have
to pay a substantial part of the disability costs for 5 years. In
order to reduce the employers’ concomitant incentive to screen
out employees with high sickness or disability risk, employers
are exonerated from this penalty if they hire a person with a
registered work handicap, and they receive reductions in their
sickness and disability insurance premia when keeping or hiring
Employment Protection Reform
24
employees with work handicaps.16
Severance payments
Severance payments mandated by French law are relatively low,
and grow more than proportionally with seniority: 2/10 months
per year of seniority, plus, for workers with more than 10 years,
2/15 months per year above 10 years. This gives 2 months for a
worker with 10 years seniority, 8.3 months for a worker with 30
years seniority.
“Quasi-legal” levels of severance pay (meaning: according to the
formula agreed upon by sub-district courts. In 1997, 75% of the
sub-district judges reported to have used this formula) are more
generous in the Netherlands. The number of months received
depends on the number of years in service and the age of the
worker. For example, a 50-year old with 10 years of service receives 15 months of severance payments. A correction factor can
also be applied (for example a layoff motivated by poor employee
behavior cuts the severance pay by anything from 0 to 100%.
Conversely, the correction factor may exceed 1 in case the employer is to blame).17 Administrative bodies do not determine
16. For more on incentives in this system, see van der Ploeg (2003).
17. The courts’ formula goes as follows: The level of severance payment is
given by the product A × B × C where
– A is a weighted average of number of years of service. Years of
service receive weight 1 before age 40, 1.5 between age 40 and age
50, and 2 after age 50;
– B is the last earned monthly wage of the employee;
– C is the correction factor.
Employment Protection Reform
25
severance payments, but they define a “noticing period”, over
which the worker still receives pay.
2.3
Procedure
France: Firms that decide to lay workers off for personal or
economic reasons must follow an often long series of administrative steps. These steps have two separate purposes. The first is
to give time to the workers to prepare themselves for the layoff and to facilitate their reemployment. Depending on seniority,
workers get an advance notice of up to three months. Workers in
large firms (1000 employees or more) are entitled to a retraining
period (“congé reclassement”) of 4 to 9 months. For the part of
the period that coincides with the advance notice period, workers get 100% of their salary; for the rest of the period, they get
65% of their salary, paid by the firm. Under the new unemployment insurance system, workers in smaller firms are eligible for
training and help in finding jobs from the start of their advance
notice (the “PARE anticipé”), not the moment they become
unemployed. The other purpose is, officially, to make sure that
alternatives to the layoffs have been fully explored.
The steps (which must take place before workers are notified of
the layoff) grow more numerous with the size of the firm, and the
size of the layoffs. For layoffs for personal reasons, the steps are
typically minimal—an interview and the sending of an official
letter. For layoffs for economic reasons, and for firms with more
than 100 workers, the process can take up to half a year. The
steps involve a number of meetings with the representatives of
Employment Protection Reform
26
the workers, the presentation by the firm of a detailed “plan to
save jobs” (“Plan de sauvegarde de l’emploi”), the approval of
the labor inspection office; they may also involve the nomination
of an auditor if requested by worker representatives, and the
recourse to an arbitrator if the workers’ representatives disagree
with the firm’s plan.
At the end of this process, the firm can start the advance notice
period, and then proceed with the layoffs. But the workers, if
they disagree, can go to court. Different courts have different
jurisdictions. In case of collective layoffs, workers or firms go
to regular tribunals (“Tribunaux d’instance” or “Tribunaux de
grande instance”.) For individual layoffs, as for most labor contract disputes, the standard court is the a labor tribunal known
as the “Prud’hommes”, an institution created in 1806. Each such
tribunal has two elected union representatives and two elected
representatives from employers’ organizations. In case of a tie,
the decisive vote is cast by a professional judge. When a case is
taken to the Prud’hommes, the first step is an attempt at arbitration (“audience de conciliation”). The second is a judgment
(“audience de jugement”), which can decide that layoffs were
not justified, and impose fines and payments to the firm. (98%
of the cases are brought by workers, only 2% by firms; 80% of the
cases are decided in favor of workers). The judgment can then
be appealed, going first to the appeals court (“Cour d’appel”),
then possibly to the highest court (“Cour de cassation”); 50%
of the cases are appealed, 70% are decided in favor of workers.
The number of cases taken by the Prud’hommes has increased
Employment Protection Reform
27
rapidly in the recent past, reaching close to 200,000 new cases
(half of those related to layoffs) per year at the end of the 1990s.
Both at regular tribunals, and at the Prud’hommes, the delays in
reaching a decision can be substantial (the mean time to the first
judgment at Prud’hommes is now around 10 months). If layoffs
are found not to be justified, the firm has to pay additional
severance payments. These payments can be substantial. If for
example the firm has more than 11 employees, and the worker
has more than two years seniority, severance payments must be
at least equal to six months.
Netherlands: The Dutch process is quite similar in spirit, with a
prior-check institution in place for nearly 60 years, but it is much
more expedient.18 Dutch firms are offered a menu of options:
The first option is the administrative process. After an advance
notice period (which in the case of individual layoffs increases
with the worker’s seniority and goes from 1 to 4 months), the
firm can attempt to obtain authorization from a public administrative body, currently the Center for Work and Income (CWI).
The procedure is rather lengthy, but cheap.
Alternatively the firm can make its case to the sub-district court
(without advance notice period). The procedure is then more expensive, but involves smaller delays. In principle, the procedure
is completed within a month. The costs of the administrative and
court procedures are estimated at about 5,000 and 27,225 euros,
18. In 1975, France introduced the need for prior administrative authorization. This requirement was eliminated in 1986.
Employment Protection Reform
28
respectively (van Zevenbergen-Oelen 2000). Despite the cost, the
swift procedure appeals to many firms, especially large ones. For
the first time, in 1998, the number of requests for dissolution filed
at sub-district courts exceeded the number of requests filed with
the administrative bodies. In either case decisions are seldom in
favor of workers. For example, for individual dismissals and the
administrative procedure, only 6% of the firms’ requests are rejected (CPB1997). According to van Zevenbergen-Oelen (2000),
the courts and administrative bodies allowed the layoffs in 96.5%
of cases. Of course, such statistics must be taken with caution;
after all 20% of the layoff demands are withdrawn and employers
do not seem to be willing to engage in a procedure unless they
have a particularly strong case.
2.4
Instruments of flexibility: fixed duration
contracts and outsourcing
a) Fixed-duration contracts
France: Since the late 1970s, successive governments have introduced fixed–term contracts, called “contrats à durée determinée”, or CDDs. These contracts still require a severance payment, but eliminate the recourse to courts when termination
takes place at the end of the contract.19
A brief history of CDDs goes as follows: CDDs were introduced
in 1979. With the election of a socialist government in 1981 and
the passage of another law in 1982, their scope was reduced: A
list of 12 conditions was drawn, and only under those conditions
19. Poulain (1994) gives a detailed description of the rules governing CDDs.
Employment Protection Reform
29
could firms use fixed-term contracts. In 1986, the 12 conditions
were replaced by a general rule: CDDs should not be used to fill
a permanent position in the firm. The current architecture dates
for the most part to an agreement signed in March 1990. Under
this agreement, CDDs can be offered by firms for only one of four
reasons: (1) The replacement of an employee on leave (2) Temporary increases in activity (3) Seasonal activities (4) Special
contracts, aimed at facilitating employment for targeted groups,
from the young to the long term unemployed. The list of special
contracts has grown in the 1990s, as each government has tried
to improve labor market outcomes for one group or another;
some of these contracts require the firm to provide training, and
many come with subsidies to firms.
Fixed duration contracts are subject to a very short trial period, typically one month. Their duration goes from 6 to 18
months, depending on the specific contract type. Mean duration
is roughly one year. They typically cannot be renewed, and, in
any case, cannot be renewed beyond 24 months. If the worker is
kept, he or she must then be hired on a regular open-ended contract, called a Contrat à durée indéterminée (CDI). If the worker
is not kept, he or she receives a severance payment equal to 10%
of the total salary received during the life of the contract.20 Note
that this is a much higher percentage of salary than is the case
for severance on regular contracts. But, as emphasized earlier,
workers on CDDs cannot go to the Prud’hommes to contest the
end of employment on the CDD.
20. As of January 2003. A collective agreement can bring this “prime de
précarité” down to 6% in exchange for additional training.
Employment Protection Reform
30
Fixed duration contracts have been very popular with firms, and
represent now 70% of the flow of hires, and a bit above 10% of
total employment.
Netherlands: By contrast, Dutch workers on fixed duration contracts receive no severance pay. However, the divide between
fixed duration and open-ended contracts does not seem as sharp
as in France. Two recent reforms, the Flexibility and Security
Act (that originated in a 1996 Labour Foundation agreement
between the social partners and came into force on January 1,
1999), and a new law in 2001, grant new rights to fixed-duration
contracts: transformation, under certain conditions, of consecutive temporary employment contracts into a permanent one,
requirement for firms to offer vacancies for open-ended contracts
to employees on fixed-duration contracts, treatment of contracts
with temporary employment agencies as permanent contracts.
As Table 2 on new work relationships demonstrates, the share of
temporary workers in the flow of new hires is much smaller than
in France (while the 2000 percentage of workers on a temporary
contract, 12%, is comparable with the French level):
Employment Protection Reform
Terms of employment
1994
Number
31
1995
%
Number
1997
%
Number
Stable
313
50%
375
50%
570
Temporary work
76
12%
105
14%
139
Remaining temporary
177
28%
199
27%
193
Freelance
40
6%
46
6%
58
Remaining
26
4%
20
3%
58
Total
632
745
1.018
Source: Research voor Beleid/Arbeidsvoorziening (HZW2000)
1998
1999
2000
%
Number
%
Number
%
Number
%
56%
14%
19%
6%
6%
661
154
226
64
25
1.130
58%
14%
20%
6%
2%
733
125
245
42
32
1.177
62%
11%
21%
4%
3%
723
99
200
39
38
1.108
65%
9%
18%
4%
3%
Table 2
The Dutch unions’ willingness to improve the rights of temporary workers in exchange of a relaxation of statutory dismissal
protection is as remarkable as their actions in favor of part-time
work and integration of women in the work force. By contrast,
French unions traditionally focus (except rhetorically) on the
interests of workers with open-ended contracts.
b) Outsourcing
In the Netherlands, the Flexibility and Security Act made it
more attractive for traditional employers to outsource, as they
can call the same workers as many times as they want (there is no
limitation to the number of contracts an employer can sign with
an employee of a temporary work agency-TWA). The employee
has a regular contract with the temporary work agency. The new
regulation improved the rights of the employees of TWAs but
do not require more from employers employing workers through
Employment Protection Reform
32
these TWA. In that sense, it provides flexibility to traditional
employers and reduces their labor costs (for example, they do
not have to pay firing costs; they can simply not re-employ the
worker at the end of the temporary contract).
3
Contours of employment protection
reform in France and in the Netherlands
In our report on employment protection reform in France, we
concluded that there was a strong case for reform along three
major dimensions: An increase in the marginal financial cost of
layoffs for firms, a decrease in the role of courts, and a reduction of the sharp contrast between permanent and temporary
contracts. It would be presomptuous for us to make specific recommendations for the Netherlands. But our assessment is that
the first dimension is equally important for the Netherlands,
while the other two are probably less important: In those two
dimensions, Netherlands has largely avoided some of the problems France is now facing.
Let us review the three dimensions in turn:
1:
An increase in the marginal financial cost of layoffs for
firms.
We saw that the contribution rate of firms should be positive,
although probably less than one.
Starting either from the French or the Dutch current legislation,
this implies a reduction in the payroll tax, and the introduction
Employment Protection Reform
33
of unemployment contributions by firms related to their layoff
behavior. While shifting to a positive contribution rate will induce firms to reduce layoffs, this increase in employment protection (with payments from firms to the unemployment insurance
agency, rather than directly to workers) will be less visible to
workers than some of the other forms of employment protection.
But it is nevertheless an increase in employment protection: It
leads firms to take into account the social costs of unemployment, and decrease their layoff rate.
If and when such a payroll tax is introduced, is there a role
left for severance payments, direct payments to workers? We
think so, but their role should be only to offset the costs of job
loss (as separate from unemployment). This should be their only
and limited purpose; unemployment insurance is better provided
through unemployment benefits. Given that the costs of job loss
appear to be increasing and convex in seniority, this suggests
the use of a schedule which is increasing and convex in seniority,
with low payments until high seniority is achieved.
Here, there may be a relevant difference between France and the
Netherlands. Legally mandated severance payments in France
are, as we saw, relatively low. Quasi-legal severance payments
in the Netherlands are significantly higher. Adding a layoff tax
to these severance payments may lead to excessive payments
by firms in case of layoff. In this case, it may make sense to
decrease severance payments paid by firms as the layoff tax is
introduced. This however will have distributional implications:
If unemployment insurance payments are not increased, workers
Employment Protection Reform
34
will receive smaller overall payments in case of unemployment
(smaller severance payments, same unemployment benefits).
In any case, if severance payments are convex in seniority, there
are constraints, as we saw in our discussion of the Delalande
contribution in France, on how steep the schedule can be at
high seniority. If it is too steep, it runs the risk of generating
discrimination against middle–age workers.
In case of bankruptcy, firms should be liable for layoff taxes and
severance payments, and the state should be a senior creditor.
As we know however from recent cases, firms have an incentive to escape those liabilities by designing complex structures
of ownership so as to benefit from limited liability. The problem will only grow more serious, if, as we argue should happen,
contribution rates are increased.
2:
A decrease in the role of courts in case of layoffs, leading
to a less costly and less uncertain process for firms.
In light of our discussion, the heavy hand of the judicial process,
as it now exists in France, seems largely unjustified. We do not
see why the Prud’hommes or tribunals should be asked to second
guess the decision of the firm, if the firm goes through the proper
administrative steps and is willing to pay both contributions to
the UI fund and severance payments to its workers.
The role of the tribunals in France should therefore be much
more limited than it is today. Courts should, if requested by the
dismissed workers or the fund, check that proper administrative
steps have been taken, and that contributions and severance are
Employment Protection Reform
35
being paid, and, in the case of an individual layoff, that the firm
did not discriminate (e.g., against a union representative or a
pregnant woman) and did not harass the worker to masquarade
a layoff as a quit.
In this respect, the Netherlands appears in much better shape
than France. In practice, courts are much more expeditious.
While one might worry that the need for administrative authorization led to a heavy burden on firms (this was the case
in France when such an authorization was in place), this does
not appear to be the case. Thus, it does not appear essential to
modify that aspect of the employment protection system in the
Netherlands.
3:
The sharp contrast between fixed-duration and open-ended
contracts should be eliminated.
The elimination of the two-contract regime in place in France
should reduce the dual nature of the labor market, which we see
as a major and perverse effect of recent reforms. The increase
in the financial marginal cost of laying off a worker, compensated by a decrease in the complexity and the uncertainty of the
layoff process, might well be more attractive both to firms and
to workers. The appeal for French firms of fixed-duration contracts, which combine a higher severance pay than open-ended
contracts, with a much simpler process of termination, suggests
that firms would be eager to accept such a trade off. But we
believe that this need not come with a decrease in the welfare
of workers, both those on fixed-duration contracts, and those
on open-ended contracts. Given its goals, the current system is
Employment Protection Reform
36
inefficient. Efficiency gains can make both sides better off.
Here, again, the Netherlands appears to have less of a problem,
and indeed, may offer some directions of reform for France. Even
if political considerations make it difficult for France to shift
from a two-contract regime to a common regime, the purpose
should be to reduce the differences between the two regimes over
time, so as to avoid the sharp threshold effects that characterize
the current system.
Employment Protection Reform
37
References
[1] Belot, M., and J. van Ours (2001) “Unemployment and Labor Market Institutions: An Empirical Analysis,” Journal
of the Japanese and International Economics ,” 15: 403–
418.
[2] Blanchard, O., and J. Tirole (2003a) “Protection de
l’Emploi et Procedures de Licenciement”, Conseil d’Analyse
Economique, Documentation Francaise, numero 44. (English version: “Contours of Employment Protection Reform”, MIT working Paper, November 2003)
[3] Blanchard, O., and J. Tirole (2003b) “The Design of Optimal Labor Market Institutions”, mimeo.
[4] Boeri, T.(2002) “Making Social Europe Compete,” mimeo,
Universita Bocconi.
[5] CFDT (2003) Salariés: Guide de vos droits 2003. La
Découverte; Guides.
[6] CPB Netherlands Bureau of Economic Policy Analysis (1997) Challenging Neighbours: Rethinking German
and Dutch Economic Institutions. Berlin and New York:
Springer.
[7] den Butter, F.A.G., and R. Mosch (2003) “The Dutch Miracle: Institutions, Networks, and Trust”, Journal of Institutional and Theoretical Economics, 159, 362–391.
[8] Fougère, D., and D. Margolis (2000), “Moduler les
cotisations employeurs à l’assurance chômage : les
expériences de bonus-malus aux Etats-Unis,” Revue
Française d’Economie, 2: 3–76.
Employment Protection Reform
38
[9] Frederiksson, P., and B. Holmlund (2003) “Improving Incentives in Unemployment Insurance: A Review of Recent
Research,” CES IFO Working Paper 922.
[10] JurisClasseur Groupe Lexis-Nexis (2002) Le licenciement
pour motif économique après la loi de modernisation sociale, sous la direction de Christophe Willmann et JeanYves Kerbough.
[11] Pélissier,J., Supiot, A., and A. Jeammaud (2002) Droit du
Travail. Dalloz, Paris.
[12] Poulain, G. (1994) Les contrats de travail à durée
déterminée. Litec, Paris, France.
[13] van der Ploeg, R. (2003) “Role of Employers in Transforming Disability into Ability: Learning from Dutch Mistakes,”
mimeo.